@inproceedings{pei-etal-2026-clore,
title = "{C}lo{R}e: Closed-Loop Reasoning via Intermediate-State Maintenance for Multi-Hop Question Answering",
author = "Pei, Yingxin and
Han, Yuantao and
Li, Ruiting and
Ji, Binquan and
Luo, Haibo and
Lu, Yifei and
Ren, Feiliang",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1404/",
pages = "29924--29953",
ISBN = "979-8-89176-462-0",
abstract = "Retrieval-Augmented Generation (RAG) has substantially advanced knowledge-intensive question answering (QA). Existing methods typically follow a `decompose and sequentially solve' paradigm, mainly optimizing sub-question generation, retrieval, or search decisions, while lacking explicit maintenance of cross-step intermediate reasoning states. As a result, retrieval noise, query drift, and local reasoning errors introduced in early steps can be inherited and amplified along the reasoning chain, eventually causing reasoning failure. We propose **CloRe**, a state-based closed-loop reasoning framework for multi-hop question answering that explicitly maintains hop-level intermediate states and coordinates forward review with dependency-aware backward correction. In the forward phase, CloRe uses conditional query rewriting to mitigate query drift, stage-wise evidence acquisition to mitigate retrieval noise, and step-level review with local repair to filter unreliable intermediate answers before they are propagated. When forward correction is insufficient, CloRe enters a backward phase that uses downstream failure signals to backtrack along explicit dependencies and revise predecessor states that were locally acceptable under earlier review but become incompatible with later dependency constraints. Experiments on three representative multi-hop QA benchmarks show that CloRe achieves competitive performance against strong baselines, demonstrating the effectiveness of explicit intermediate-state maintenance for multi-hop reasoning."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="pei-etal-2026-clore">
<titleInfo>
<title>CloRe: Closed-Loop Reasoning via Intermediate-State Maintenance for Multi-Hop Question Answering</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yingxin</namePart>
<namePart type="family">Pei</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuantao</namePart>
<namePart type="family">Han</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruiting</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Binquan</namePart>
<namePart type="family">Ji</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haibo</namePart>
<namePart type="family">Luo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yifei</namePart>
<namePart type="family">Lu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Feiliang</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Retrieval-Augmented Generation (RAG) has substantially advanced knowledge-intensive question answering (QA). Existing methods typically follow a ‘decompose and sequentially solve’ paradigm, mainly optimizing sub-question generation, retrieval, or search decisions, while lacking explicit maintenance of cross-step intermediate reasoning states. As a result, retrieval noise, query drift, and local reasoning errors introduced in early steps can be inherited and amplified along the reasoning chain, eventually causing reasoning failure. We propose **CloRe**, a state-based closed-loop reasoning framework for multi-hop question answering that explicitly maintains hop-level intermediate states and coordinates forward review with dependency-aware backward correction. In the forward phase, CloRe uses conditional query rewriting to mitigate query drift, stage-wise evidence acquisition to mitigate retrieval noise, and step-level review with local repair to filter unreliable intermediate answers before they are propagated. When forward correction is insufficient, CloRe enters a backward phase that uses downstream failure signals to backtrack along explicit dependencies and revise predecessor states that were locally acceptable under earlier review but become incompatible with later dependency constraints. Experiments on three representative multi-hop QA benchmarks show that CloRe achieves competitive performance against strong baselines, demonstrating the effectiveness of explicit intermediate-state maintenance for multi-hop reasoning.</abstract>
<identifier type="citekey">pei-etal-2026-clore</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1404/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>29924</start>
<end>29953</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T CloRe: Closed-Loop Reasoning via Intermediate-State Maintenance for Multi-Hop Question Answering
%A Pei, Yingxin
%A Han, Yuantao
%A Li, Ruiting
%A Ji, Binquan
%A Luo, Haibo
%A Lu, Yifei
%A Ren, Feiliang
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F pei-etal-2026-clore
%X Retrieval-Augmented Generation (RAG) has substantially advanced knowledge-intensive question answering (QA). Existing methods typically follow a ‘decompose and sequentially solve’ paradigm, mainly optimizing sub-question generation, retrieval, or search decisions, while lacking explicit maintenance of cross-step intermediate reasoning states. As a result, retrieval noise, query drift, and local reasoning errors introduced in early steps can be inherited and amplified along the reasoning chain, eventually causing reasoning failure. We propose **CloRe**, a state-based closed-loop reasoning framework for multi-hop question answering that explicitly maintains hop-level intermediate states and coordinates forward review with dependency-aware backward correction. In the forward phase, CloRe uses conditional query rewriting to mitigate query drift, stage-wise evidence acquisition to mitigate retrieval noise, and step-level review with local repair to filter unreliable intermediate answers before they are propagated. When forward correction is insufficient, CloRe enters a backward phase that uses downstream failure signals to backtrack along explicit dependencies and revise predecessor states that were locally acceptable under earlier review but become incompatible with later dependency constraints. Experiments on three representative multi-hop QA benchmarks show that CloRe achieves competitive performance against strong baselines, demonstrating the effectiveness of explicit intermediate-state maintenance for multi-hop reasoning.
%U https://aclanthology.org/2026.findings-emnlp.1404/
%P 29924-29953
Markdown (Informal)
[CloRe: Closed-Loop Reasoning via Intermediate-State Maintenance for Multi-Hop Question Answering](https://aclanthology.org/2026.findings-emnlp.1404/) (Pei et al., Findings 2026)
ACL